Hamiltonian formulation 19
=−
∑N
i=1
p^2 i
2 mi
−U(r 1 ,...,rN). (1.6.2)
The function−L ̃(r 1 ,...,rN,p 1 ,...,pN) is known as theHamiltonianH:
H(r 1 ,...,rN,p 1 ,...,pN) =
∑N
i=1
p^2 i
2 mi
+U(r 1 ,...,rN). (1.6.3)
The Hamiltonian is simply the total energy of the system expressed as a function of
positions and momenta and is related to the Lagrangian by
H(r 1 ,...,rN,p 1 ,...,pN) =
∑N
i=1
pi·r ̇i(pi)−L(r 1 ,...,rN,r ̇ 1 (p 1 ),....,r ̇N(pN)). (1.6.4)
The momenta given in eqn. (1.6.1) are referred to asconjugateto the positions
r 1 ,...,rN.
The relations derived above also hold for a set of generalized coordinates. The
momentap 1 ,...,p 3 Nconjugate to a set of generalized coordinatesq 1 ,...,q 3 Nare given
by
pα=
∂L
∂q ̇α
, (1.6.5)
and the Hamiltonian becomes
H(q 1 ,...,q 3 N,p 1 ,...,p 3 N) =
∑^3 N
α=1
pαq ̇α(p 1 ,...,p 3 N)
−L(q 1 ,...,q 3 N,q ̇ 1 (p 1 ,...,p 3 N),...,q ̇ 3 N(p 1 ,...,p 3 N)). (1.6.6)
Now, according to eqn. (1.4.18), sinceGαβis a symmetric matrix, the generalized
conjugate momenta are
pα=
∑^3 N
β=1
Gαβ(q 1 ,...,q 3 N) ̇qβ. (1.6.7)
Inverting this, we obtain the generalized velocities as
q ̇α=
∑^3 N
β=1
G−αβ^1 (q 1 ,...,q 3 N)pβ, (1.6.8)
where the inverse of the mass-metric tensor is
G−αβ^1 (q 1 ,...,q 3 N) =
∑N
i=1
1
mi
(
∂qα
∂ri
)
·
(
∂qβ
∂ri
)
. (1.6.9)
It follows that the Hamiltonian in terms of a set of generalized coordinates is
H(q 1 ,...,q 3 N,p 1 ,...,p 3 N) =
1
2
∑^3 N
α=1
∑^3 N
β=1
pαG−αβ^1 (q 1 ,...,q 3 N)pβ